Aspiring AI Engineer with Full-Stack Focus

YAJUR KHANNA

Boston, MA, ***** 857-***-**** https://yajur-khanna.github.io/Yajur-Website/ ******.**@************.*** Available: Jan. – Aug 2026

EDUCATION

Sept. 2025 – Present

Expected Graduation: 2027

Sept. 2021 – May. 2025

Khoury College of Computer Sciences, Northeastern University, Boston, MA Master of Science in Artificial Intelligence - Candidate

Related courses: Foundations of Artificial Intelligence, Algorithms, Machine Learning Vellore Institute of Technology, Chennai, India

Bachelor of Engineering in Computer Science,

Related courses: Real-Time Operating Systems, Databases, Embedded Systems TECHNICAL KNOWLEDGE

Languages:

Skills:

Libraries:

Certifications:

Python, JS, HTML, CSS, R, MySQL, Go

Computer Vision, RL, NLP, Machine Learning, AWS(S3, Lambda)/Azure, Deep Learning, Flask, RAG, Github, HuggingFace, CI/CD Pipelines, FastAPI, Docker, MongoDB, ONNX PyTorch, Keras, TensorFlow, OpenCV, Kubernetes, Pandas, NumPy, Kafka, LangChain, Redis Big Data Analysis using DL and AWS (National University of Singapore, Score - 77%) WORK EXPERIENCE

Infotrixs, Mumbai, India

Python Developer Intern Oct 2023 – Nov 2023

• Developed Python backend service leveraging WeatherAPI REST API integration to fetch and parse real-time JSON weather data with automatic refresh capabilities every 30 minutes and comprehensive exception handling

• Built persistent storage backend with file I/O operations for managing user favorite cities, implementing data serialization, retrieval logic, and API endpoints for frontend consumption, reducing data access latency by 40% Corporate Gurukul, National University of Singapore, Singapore Teaching Assistant May 2025 – Jul 2025

• Delivered lectures on AI fundamentals (supervised/unsupervised learning, neural networks) to 25 high school students, achieving 95% positive feedback

• Mentored student groups in developing AI chatbots using AWS services (Lambda, Lex), providing hands-on guidance on NLP implementation, intent recognition, and cloud deployment, resulting in 6 deployed chatbot prototypes CybTree, Chennai, India

AI Automation Intern Feb 2025 – Apr 2025

• Developed an automated Attack Surface Management (ASM) tool using Python and Typer CLI framework for domain and subdomain security analysis reducing manual security analysis time by 60%

• Designed a risk scoring engine using NLP to quantify security posture and generated AI-driven summary of findings

• Built JSON reporting pipeline compiling 10 findings per domain and cutting documentation time by 40% PROJECTS

Real Time Factory Defect Detection System with Scalable ML Deployment Oct 2025 – Dec 2025

• Built real-time defect detection pipeline using MVTec AD, deploying CNN model via FastAPI on Kubernetes with Kafka streaming and sub-60ms Redis-cached inference

• Added explainable outputs with calibrated confidence and defect sizing, improving defect localization accuracy by 12% and reducing low-confidence prediction rates by 22% Skin Cancer Detection using Hybrid Deep Learning Models and HAM10000 Dataset Dec 2023 – May 2024

• Built a dual-track Swin-UNet + MobileNetV3 model for HAM10000 achieving 83% accuracy on 7-class skin-lesion classification. Deployed the model in a web app enabling AI assisted diagnosis of skin lesions via image uploads Flask-Based Adaptive AI Interview Platform with GPT-4 Integration Dec 2024 – Apr 2025

• Created an AI-powered interview platform using Flask, SocketIO, and GPT-4 for dynamic question generation with sentence-transformer based deduplication (cosine similarity < 0.65) reducing similar question generation by 85%

• Architected 5-round progressive interview system with adaptive follow-up generation, real-time answer tracking, and configurable time management across difficulty levels, and auto-grading user answers using GPT-4 Custom Neural Network Framework with Autograd Jun 2024 – Aug 2024

• Built neural network framework from scratch with automatic differentiation via dynamic computational graph, implementing backpropagation, gradient descent optimization, and MLP training on MSE loss; validated against PyTorch's autograd and visualized computational graphs using Graphviz INTERESTS/ACTIVITIES

• Awarded Best Paper for a MobileNetV3–Swin-Unet hybrid achieving 8% higher accuracy compared to state-of-the-art models in plant leaf disease detection and a 35% reduction in model size through an optimized attention design

Contact this candidate